Skip to content

ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning & ReCall: Learning to Reason with Tool Call for LLMs via Reinforcement Learning

active 2025-04-232026-07-25 (UTC)

Partial coverage11,254 / 11,982 hourly files (94%) · 2 absent upstream · 726 failed, retryable2025-03-292026-08-10 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
416
Pushes
0
Pull requests
0
Issues
35
Stars
322
Forks
19

Activity over time

Daily event counts in the loaded window

Line chart, 459 days from 2025-04-23 to 2026-07-25. Pushes: 0 total, peak 0 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 35 total, peak 5 in a day. Comments: 40 total, peak 5 in a day. Stars: 322 total, peak 14 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

Recent activity

Latest issues, pull requests and releases

  • Issue comment#91Hyfred2026-01-23 05:46
    Released model Performance lower than reported in the paper
  • Issue#91Hyfred2026-01-21 02:10
    Released model Performance lower than reported in the paper
  • Issue#88callmeluoq-art2025-08-28 08:39
    How to support asynchronous tool calls?
  • Issue comment#53cwszz2025-08-21 07:44
    Retriever Serving error
  • Issue#87threegold1162025-08-21 02:58
    model merge
  • Issue#84songjiechong2025-08-14 09:28
    数据集相关问题
  • Issue comment#60BaiMeiyingxue2025-08-14 01:28
    OOV Train Error
  • Issue comment#83Alethes02162025-07-31 08:26
    Guidance on Correctly Enabling Flash Attention 2 with Qwen2-7B-Instruct (dtype and device warnings)
  • Issue comment#81Alethes02162025-07-25 08:25
    value error: Token id 151908 is out of vocabulary
  • Issue comment#81YurainSoon2025-07-23 09:27
    value error: Token id 151908 is out of vocabulary
  • Issue comment#80AnselCmy2025-07-23 08:32
    Mismatch between loss_mask and attention_mask/seq length — will this cause issues during training?
  • Issue#81Alethes02162025-07-23 04:57
    value error: Token id 151908 is out of vocabulary
  • Issue comment#30BaiMeiyingxue2025-07-22 09:00
    datasets.arrow_writer.SchemaInferenceError: Please pass `features` or at least one example when writing data
  • Issue#79songjiechong2025-06-30 06:47
    训练过程reward score问题
  • Issue#79songjiechong2025-06-29 10:42
    训练过程reward score问题
  • Issue#78qilong-zhang2025-06-20 10:01
    exchange wechat
  • Issue#77prasadke202025-06-14 12:58
    Requesting compatibility for LORA based fine-tuning with GRPO
  • Issue comment#76prasadke202025-06-13 09:11
    Reward Not Increasing While trying to use Qwen 2.5 - 0.5B,1.5B Instruct models for training on musique with re-search code
  • Issue#75AnselCmy2025-06-13 06:33
    QwQ-32B with Tool Use as Baseline?
  • Issue#38AnselCmy2025-06-13 06:32
    Training with an ERROR, and speed issue
  • Issue#45AnselCmy2025-06-13 06:32
    About preprocessed training data
  • Issue#49AnselCmy2025-06-13 06:32
    Why setting the max tokens for each turn to 512?
  • Issue comment#76AnselCmy2025-06-13 06:19
    Reward Not Increasing While trying to use Qwen 2.5 - 0.5B,1.5B Instruct models for training on musique with re-search code
  • Issue comment#76prasadke202025-06-12 19:47
    Reward Not Increasing While trying to use Qwen 2.5 - 0.5B,1.5B Instruct models for training on musique with re-search code
  • Issue#76prasadke202025-06-12 19:35
    Reward Not Increasing While trying to use Qwen 2.5 - 0.5B,1.5B Instruct models for training on musique with re-search code

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 322 stars here means stars gained during the window, not the repo's star count.